September1999

Writing Style

Excerpt from notes by

Prof. Marie desJardins

September19992

Sources

Robert L. Peters, Getting What You Came For: The Smart Student’s Guide to Earning a Master’s or Ph.D. (Revised Edition). NY: Farrar, Straus, and Giroux, 1997.

Justin Zobel, Writing for Computer Science: The Art of Effective Communication, 2/e. London: Springer-Verlag, 2004.

Also useful: Lyn Dupré, BUGS in Writing. Addison Wesley, 1995.

September19993

Questions

How many of you like to write [in English]? How many of you think you’re good at writing [in

English]? How many of you are worried about writing [for this

class, for your thesis/dissertation]?

September1999

Technical Writing Style

September19995

Robert’s Words of Wisdom

Keep it brief. Break it up. Don’t be self-important. Start your paragraphs with topic sentences. Don’t write a detective novel. Don’t try to handle too many ideas at once. Use key words. Signpost with transitional phrases. Repeatedly summarize. Avoid passive constructions. Avoid adverbs. Delete double negatives. Chop off your first paragraph. Read it out loud. Read it again cold. Move back and forth between word processor and paper.

Quoted from Peters pp. 231─233

September19996

Zobel in a Nutshell

Simplicity is key.

September19997

Don’t Be Obscure

Example from Peters (p. 229/230): “[The Environmental Protection Agency] has developed an industry-

specific cross-media pollution-abatement model that also estimates the reduction in human health risks attributable to adopting various sets of abatement measures. The model has been applied to the iron and steel industry.”

Rewrite: “In order to understand how to reduce pollution in some specific

industries, the [EPA] has developed a computer model which examines how pollutants in air, water, and other environmental media interact. In addition, the model can estimate how selected measures to reduce pollution would also reduce human health risks. As a trial run, the EPA has used this model to examine pollution reduction in the iron and steel industry.”

September19998

Cut It Out I

The volume of information has been rapidly increasing in the past few decades. While computer technology has played a significant role in encouraging the information growth, the latter has also had a great impact on the evolution of computer technology in processing data throughout the years. Historically, many different kinds of databases have been developed to handle information, including the early hierarchical and network models, the relational model, as well as the latest object-oriented and deductive databases. However, no matter how much these databases have improved, they still have their deficiencies. Much information is in textual format. This unstructured style of data, in contrast to the old structured record format data, cannot be managed properly by the traditional database models. Furthermore, since so much information is available, storage and indexing are not the only problems. We need to ensure that relevant information can be obtained upon querying the database. (Zobel p. 12)

September19999

Cut It Out I

The volume of information has been rapidly increasing in the past few decades. While computer technology has played a significant role in encouraging the information growth, the latter has also had a great impact on the evolution of computer technology in processing data throughout the years. Historically, many different kinds of databases have been developed to handle information, including the early hierarchical and network models, the relational model, as well as the latest object-oriented and deductive databases. However, no matter how much these databases have improved, they still have their deficiencies. Much information is in textual format. This unstructured style of data, in contrast to the old structured record format data, cannot be managed properly by the traditional database models. Furthermore, since so much information is available, storage and indexing are not the only problems. We need to ensure that relevant information can be obtained upon querying the database. (Zobel p. 12)

September199910

Cut It Out I Redux

Much information is textual. This unstructured data cannot be managed properly by traditional database models. Furthermore, storage and indexing are not the only problems. We need to ensure that relevant information can be obtained upon querying. (Zobel p. 12)

Can you do better?

September199911

Cut It Out I Redux

Much information is Unstructured textual. This unstructured data, cannot be managed properly by traditional database models. Furthermore, Storage and indexing are not the only problems.: we also need to ensure that relevant information can be obtained upon querying.

September199912

Cut it Out II

As part of their work, they showed that the problem of finding the best total order in a set of given items belongs to the class of NP-hard problems. To be able to find an approximation for the global order, in the paper, they provide a simple greedy algorithm and a second slightly modified algorithm that takes advantage of strongly connected graphs to return an approximation to the best ordering.(from a student’s paper summary of Cohen et al.)

September199913

Cut it Out II

As part of their work, they The authors showed that the problem of finding the best total order in a set of given items belongs to the class of is NP-hard problems. To be able to find an approximation for the global order, in the paper, They provide a simple two variations of a greedy algorithm and a second slightly modified algorithm that takes advantage of strongly connected graphs to return an approximateion to the best ordering.

September199914

Cut it Out III

The contribution of this paper was to describe a method in which a collection of objects can be ordered, using preference judgments. There are two stages in which this ordering is done. First, one creates a “binary preference function” in order to determine how to rank the objects. Then, one uses this function to order the objects.(from a student’s paper summary of Cohen et al.)

September199915

Cut it Out III

The contribution of this paper is was to describe a two-stage method in which a for ordering a collection of objects can be ordered, using preference judgments. There are two stages in which this ordering is done. First, one creates In this approach, a “binary preference function” is learned from training data; in order to determine how to rank the objects. then, one uses this function is used to order the a new set of objects.(from a student’s paper summary of Cohen et al.)

September199916

Writing with Clarity

Don’t write overly long papers, sections, paragraphs, sentences, or words

Know what each section, paragraph, and sentence is about, and stick to the subject

Define your terms, and use boldface or another convention to make them stand out

Expand your acronyms (and use as few as possible)

Explain your math in English

September199917

Responding to Criticism

“The reader is always (well, at least sometimes) right (or at least kinda).”

Don’t get defensive and start making excuses: “It’s in there!” [Then why didn’t they notice it?] “I didn’t have room!” [Then maybe you should rethink

your priorities.] “It’s not important!” [But this reader thinks it is. So the

paper has to explain it, or convince her that it’s not important.]

But ignore “There’s no future work” comments...*

*Anecdote alert!

September199918

Revisiting Paraphrasing

Be careful about paraphrasing: Zobel p. 25-26: According to Fier and Byke such an approach is “simple

and...fast, [but] fairly crude and... could be improved”

is revised to Fier and Byke describe the approach as simple and fast,

but fairly crude and open to improvement.

Why skirt the edge of plagiarism when you can use your own words and clarify what is meant: According to Fier and Byke, this approach is efficient, but

the quality of the results could be improved.”

September199919

On Self-Plagiarism

OK under certain circumstances... Most researchers reuse parts of earlier papers, especially

related work and terminology New publications should be substantially different and/or

have significant new results

September1999

Specifics

September199921

Avoid Slang and Idioms

Zobel: “crop up,” “lose track,” “it turned out that,” “play up,” “right out,” “run the gamut,” “teased into”

Also: “lots,” “a lot,” “write up” Summaries: “good job,” “come up with,” “it was

odd,” “a difficult read” Avoid contractions (considered too informal for

science writing) Not common in American English: “viz.,” “the

works,” “hence”

September199922

Avoid Qualifiers and Adverbs

“Very,” “rather,” “simply,” “possibly,” “of course,” “naturally,” “obviously,” “just,” “pretty,” “pretty much,” “more of,” “extremely,” “seriously, “indeed,” “really”

Particularly avoid qualifying nonqualifiable words such as “unique,” “intractable,” “optimal,” and “infinite”

Avoid personalizing your remarks: Minimize the use of “I think,” “I feel,” “I believe,” “It seems”

September199923

Avoid Fluff

Zobel p. 55 – some of my favorites: adding together → adding cancel out → cancel during the course of → during for the purpose of → for in view of the fact → given the vast majority → most a number of → several whether or not → whether it can be seen that → it is a fact that → is something that can can involves (active verb) In the paper, ... They noticed that...

September199924

Parallel Construction

“I like to go swimming, riding bicycles, and I read a lot.”

“The complexity increased both in time and space.”

“The three most important things to remember are:1. Write a little every day.2. You should proofread everything before showing it to

your advisor.3. Careful of bad grammar!”

September199925

Parallel Construction II

The key findings are: The algorithm to learn the preference function, based on the Freund

and Schapire “Hedge” algorithm. An algorithm to find the ordering... is NP-complete; however, .... A system (composed of these two algorithms) to compile the result

sets of various searches often performs better than a domain expert entity.

The key contributions are: An algorithm to learn the preference function, based on Freund and

Schapire’s Hedge algorithm. A proof that finding the optimal ordering is NP-complete, and a

greedy algorithm that is guaranteed... A system (composed...) that compiles the result sets of multiple

searches, and often performs better than any individual search.

September199926

Nonsexist Writing

To avoid the use of “he,” you can: Use the plural Rewrite to avoid pronouns Name people in examples (with alternating male/female

names) OK these days to use “they” for singular nouns

September199927

Proper References

Smith, R. S. (1992). The best paper ever written. Journal of Impressive Results, vol([#]), 1-101.

Jones, P. Q. (2004). A few of my favorite algorithms. NY: Trivia Press.

Kim, A. B. (1999). Towards a framework for improved performance of high-density algorithms in dynamic domains. Proceedings of the Twentieth International Conference on Nothing in Particular (pp. 27-28). Los Angeles: Nothing Much Press.

September199928

Some of My Personal Nits

Its vs. it’s Which vs. that

“which” qualifies; “that” defines. Heuristic: Use “that” by default. http://www.kentlaw.edu/academics/lrw/grinker/LwtaThat_Versus_Which.htm

Between vs. among Dangling “this” references Affect vs. effect Continual vs. continuous Optimize vs. improve Plurals and apostrophes Colons, semicolons, and dashes i.e. / e.g. / etc. / et al. Hyphenate compound adjectives, not adverbs or nouns! Commas!

September199929

Remarks From Summaries

Use consistent tense Generally the present tense

Punctuation goes inside quotation marks “Scare quotes” vs italics to introduce new terms

Only do this the first time you use the new term

Avoid passive voice (usually) Authors and algorithms do things; they don’t just happen

September199930

Remarks From Summaries II

The summary itself should be written in a formal, “scientific paper” style The wrapup (discussion of presentation etc.) can be

more informal

“Via” “using” or “by” Semicolon vs. colon vs. comma “As” “because” or “since”